ARCHIVE
Article Archive
Literature reading, research, and engineering records.
- Reproducible Debugging: Git, Downloads, ABI, and Evidence States
- GPU Software Stacks and Correct PyTorch/JAX Timing
- Python Engineering: pyproject, uv Locking, and Build Isolation
- Paired Direct/Agent Evaluation and Progress Verification
- OpenPI × LIBERO: Serving, Observations, and Action Contracts
- VLA Execution: Action Chunks, Latency, and Feedback
- Navigation Baselines and Migration: Auditing Papers, Code, and Assets
- From Video to Simulation: Geometry, Scale, Contact, and Navigation
- Interfaces for Crowd Behavior, Body Motion, and Physical Control
- ROS 2 and Simulation: Clocks, TF, QoS, and Stale Actions
- Social Navigation Benchmarks: Contracts and Metrics
- Humanoid Robots: Control Dimensions, Symmetry, and Curriculum Recovery
- Simulation Timing, NavMesh Updates, and Profiling
- Minimal Experiments for Contact and Mecanum Wheels
- Robot Assets: CAD, USD, URDF, and Inertial Validation
- RL Implementation: Objectives, Masks, and Gradient Paths
- Autograd, Gradient Accumulation, and Training Recovery
- PyTorch: Axes, Broadcasting, and Storage Layout
- Reinforcement Learning: An Introduction — A Reading Route
- Potential-Based Shaping and Q-Value Initialization are Equivalent
- Policy Gradient Methods for Reinforcement Learning with Function Approximation
- High-Dimensional Continuous Control Using Generalized Advantage Estimation
- For SALE: State-Action Representation Learning for Deep Reinforcement Learning
- Fast Gradient-Descent Methods for Temporal-Difference Learning with Linear Function Approximation
- Era of Experience: Notes on Experience-Driven Learning
- Deep Reinforcement Learning that Matters: Supplementary Evidence
- A Closer Look at Deep Policy Gradients
- A Reproducible RL Evaluation and Reporting Protocol
- Multi-Agent RL Benchmarks and Cooperative Evaluation
- Single-Agent RL Benchmarks and Metrics
- RL Statistics, Confidence Intervals, and Tuning Budgets
- TD7 Policy Checkpoints: Separating Training and Evaluation
- Embodied RL, LLMs/VLAs, and Research Design
- Simulators, World Models, and Planning
- Meta-RL: Task Distributions and Rapid Adaptation
- Markov States, Partial Observations, and Recurrent Memory
- IPPO, COMA, and Multi-Agent Gradient Variance
- Goal-Conditioned RL, UVFA, and HER
- MDP Homomorphisms and State–Action Abstraction
- Options, SMDPs, and Intra-Option Learning
- Potential-Based Reward Shaping
- Maximum-Entropy RL, SAC, and Dual Temperature Updates
- Reparameterization and Pathwise Gradients
- Offline RL and Value Extrapolation Error
- Diffusion Policy: Distribution Modeling and Control Timing
- Behavior Cloning, Action Bins, and Multimodal Actions
- Eligibility Traces and TD(λ)
- Policy Gradients, REINFORCE, and Surrogate Objectives
- On-Policy Sampling, Off-Policy Learning, and Replay
- Contractions, Fixed Points, and Stochastic Approximation
- Probability Measures, Empirical Measures, and RL Occupancy
- MDPs, Value Functions, and Bellman Errors
- Cross-Paper Comparison and Research Roadmap
- Vision Language Action Models in Robotic Manipulation: A Systematic Review
- A Survey on Vision-Language-Action Models: An Action Tokenization Perspective
- Survey of Vision-Language-Action Models for Embodied Manipulation
- A Survey on Reinforcement Learning of Vision-Language-Action Models for Robotic Manipulation
- SimBa: Simplicity Bias for Scaling Up Parameters in Deep Reinforcement Learning
- Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware
- Diffusion Policy: Visuomotor Policy Learning via Action Diffusion
- π₀: A Vision-Language-Action Flow Model for General Robot Control
- MetaVLA: Unified Meta Co-training For Efficient Embodied Adaption
- RL²: Fast Reinforcement Learning via Slow Reinforcement Learning
- RL Token: Bootstrapping Online RL with Vision-Language-Action Models
- π_RL: Online RL Fine-tuning for Flow-based Vision-Language-Action Models
- What Can RL Bring to VLA Generalization? An Empirical Study
- Realtime-VLA FLASH: Speculative Inference Framework for Diffusion-based VLAs
- Action-aware Dynamic Pruning for Efficient Vision-Language-Action Manipulation
- Real-Time Execution of Action Chunking Flow Policies
- VLASH: Real-Time VLAs via Future-State-Aware Asynchronous Inference
- DynamicVLA: A Vision-Language-Action Model for Dynamic Object Manipulation
- Running VLAs at Real-time Speed
- Towards Generalizable Robotic Manipulation in Dynamic Environments
- MotionBricks: Scalable Real-Time Motions with Modular Latent Generative Model and Smart Primitives
- Uni-Inter: Unifying 3D Human Motion Synthesis Across Diverse Interaction Contexts
- GRAIL: Generating Humanoid Loco-Manipulation from 3D Assets and Video Priors
- NavDP: Learning Sim-to-Real Navigation Diffusion Policy with Privileged Information Guidance
- FLUX: Accelerating Cross-Embodiment Generative Navigation Policies via Rectified Flow and Static-to-Dynamic Learning
- NavIsaacLab: Generating Realistic Crowd via Parallel Robot Learning for Benchmarking Human-aware Navigation
- Arena-Bench 2.0: A Comprehensive Benchmark of Social Navigation Approaches in Collaborative Environments
- Social robot navigation: a review and benchmarking of learning-based methods
- Towards Physically Executable 3D Gaussian for Embodied Navigation
- Lyra 2.0: Explorable Generative 3D Worlds
- Hexo and GitHub Pages: Building and Publishing While Preserving Sources